Deer Vehicle Collisions in Scotland Monitoring ... - The Deer Initiative

Research Report 10/1

Pilot study to assess the potential of selected existing structures on the A30 and A38 trunk roads to provide safer crossing places for deer. J Langbein

July 2010

1

Pilot study to assess the potential of selected existing structures on the A30 and A38 trunk roads in Southwest England to provide safer crossing places for deer.

Prepared by Jochen Langbein Langbein Wildlife Associates

Deer Initiative Research Report 10/1 July 2010

Further information can be obtained from: The Deer Initiative The Carriage House Brynkinalt Business Centre Chirk Wrexham LL14 5NS www.thedeerinitiative.co.uk Or email: [email protected]

This work was commissioned by the Deer Initiative for the Highways Agency. The views contained in this report are those of the author and do not necessarily reflect those of the Deer Initiative or the Highways Agency.

Contents Page Contents

(i)

Executive Summary

(ii)

1. Introduction

1

2. Previous research into use by deer of crossing structures

4

3. Approach and Methods

9

Study site selection

9

Field survey of structures

10

Analysis and interpretation

11

4. Results and Interpretation

15

5. Discussion and further work

25

Acknowledgements

32

References

33

Appendices : Appendix I

Survey Form

Appendix II

Tables of recorded structure characteristics

Appendix III

Small size photos of structures

Appendix IV

Report CD (& including video clips, high resolution photographs, and other supplementary files)

i

Executive Summary i.

Collisions of motor vehicles with deer have escalated over the past five decades in most countries across Europe including Britain. Recent studies supported by the Highways Agency indicate that the annual toll of deer vehicle collisions (DVCs) in Britain is now very likely to exceed 42,500 and may be as high as 74,000, and lead to several hundred human injuries and a number of human fatalities each year. Although DVCs are also a serious problem in Scotland, within the UK close to 80% are recorded in England where traffic volumes are close to nine fold greater. The numbers of DVCs recorded on the strategic network of motorways and trunk roads in England managed by the Highways Agency (which make up only 2% of all roads in England) now commonly exceed 1100 per annum with many more likely to remain unreported. Such actual collisions, as well as the need to deal with many additional incidents with live or dead deer on trunk road carriageways, have serious implications for road safety and traffic delays.

ii. The only well proven method of reducing deer collisions on motorways and major trunk roads is the use of high roadside fencing, and it is most successful where it channels animals to safer crossing points. Wide “green bridges” or “landscape bridges” constructed primarily to reduce habitat fragmentation caused by infrastructure can help also to reduce animal collisions, but their high cost tends to rule them out from widespread use for DVC mitigation. To some extent the recent focus on high profile green bridge projects in continental Europe has led to a widespread misconception that much smaller or joint use structures have little potential as safer crossings for deer and other large mammals. Deer are, however, known to make (increasing) use of trunk road accommodation structures of relatively modest size not specifically built for wildlife. iii. Few previous studies have specifically investigated the use by deer of existing accommodation structures on trunk roads in England. This pilot survey was initiated to:  review past research into deer use of underpasses and overpasses and the dimensions and other features thought to influence their suitability as wildlife passages.  survey a range of existing structures on trunk roads (the study used the A38 and A30 trunk roads in Devon and Cornwall) with the aim of assessing which were likely to be used by deer or offer good potential for adaptation to encourage use by deer. iv. Several reviews from continental Europe have recommended minimum dimensions and other physical criteria of structures that increase the likelihood of use by deer. Common recommendations include that underpasses should be at least 4 m high (for roe deer, (higher for larger deer), over 7 m wide for overpasses, and have natural substrate for part of their width. Some consensus exists that how inviting a crossing structure will appear to a deer depends more on its relative openness (width x height / length) than any individual dimension. Structures that have been found to be used by deer vary widely in terms of substrate, location, joint use by motorised traffic and other disturbance, and the relative importance of such features remains difficult to define. v. Several studies in continental Europe have concluded that regular use by deer of structures open to normal road traffic is unlikely, yet recent observations have shown that deer even use road bridges without any natural substrate in some parts of England (Langbein, 2008). These findings suggest that acceptance of man-made structures by deer may differ widely between countries and regions and is likely to alter over time depending on the degree to which local deer populations become accustomed to built up areas and human habitation. It is timely therefore to assess to what extent criteria for wildlife structures established elsewhere are likely to be transferrable to the current situation in England, and to explore ii

the potential for differing types of existing structures to be adapted to enhance their use by deer and form a greater part of DVC mitigation strategies in future. vi. The field survey for the present pilot study encompassed 47 existing structures (19 overpasses / 28 underpasses) across the A30 and A38 in Devon and Cornwall (Figure 2), with road sections along these routes selected to sample locations known to be frequented by red, fallow and roe deer. For most structures survey was limited to a one-off visit to record various physical features (dimensions / substrate / juxtaposition in relation to differing habitats) and searches for indirect recent signs of deer (dung / hoof prints / hair) nearby. For three structures where some clear signs of recent deer activity were noted CCTV video recording equipment was deployed for one to three nights, with the aim of obtaining some additional insights into the behaviour of deer at structures. Survey information was evaluated against criteria reported in the literature for structures known to be used by deer to help assess their potential as wildlife crossings. vii. Evidence of some deer activity within 50 m was noted for 18 of the 47 structures surveyed, but evidence that deer had actually crossed recently was only found for two structures during the one-off site visits. None of the structures surveyed occurred in sections of road fitted with deer fencing. However, evaluation of other features recorded for each structure indicated that well over half of those visited meet or exceed the generally accepted minimum size and openness criteria taken from the literature on wildlife passages (e.g. minimum heights of 4 m for underpasses, and openness indices no less than 0.75 for roe deer; see Section 2). As such many existing structures would seem to have a reasonable basis for adaptation as safer crossings for use by deer. viii. Many of the structures surveyed would not currently meet various other criteria commonly recommended for wildlife passages, in particular with respect to provision of a natural substrate or separation from joint usage by general public road traffic. However, recent observations of deer use of structures in other parts of England (e.g. Langbein, 2008) indicate that use of structures by light or intermittent public road traffic, and the presence of a hard substrate of tarmac or concrete, may be less of a hindrance for deer than previously thought. A negative feature noted for many underpasses surveyed was that side roads or tracks leading underneath trunk roads often have additional boundary fencing or thick hedges running along either side, which are not easily crossed by deer and may actually serve to divert them away unless some provision is made to improve direct accessibility. ix. Several CCTV video clips recorded during the present study at one underpass on the A38, showed fallow deer, not moving through that structure itself, but walking across the top along the trunk road verge. The footage serves to demonstrate that deer often feed calmly within just 2 to 6 m of passing trunk road traffic, and that without any fencing to lead them into the structure deer may be more likely to cross the main trunk carriageways. x. Overall, the study concluded that for a very high proportion of the existing structures inspected, appropriate low-cost adaptations are likely to encourage use by deer. To help demonstrate the feasibility and effectiveness of adaptation of existing structures as a means of DVC prevention, a number of practical demonstration projects are proposed at suitable structures identified during the present study. xi. One of the demonstration proposals concerns adaptation of either one or two underpasses located within the longstanding fallow deer DVC hotspot at Haldon Hill (A38), while the second concerns two overpasses across the A30 at woodland near Ebsworthy frequented by red deer as well as roe deer. Adaptations of these structures to enhance use by deer are likely to be achievable at low cost, ranging from simple removal of obstacles that currently prevent or reduce access by deer (likely to be achievable at costs below 1K), to provision of iii

more natural substrate for part of their width (also likely to be achievable at low cost), and provision of short lengths (from around 100 m) of lead-in fencing on one or more sides of each structure (with indicative project costs from around 5 K to 20 K). In addition, some parallel monitoring work to determine effectiveness of any physical adaptations undertaken is recommended, which ideally should include intermittent use of CCTV surveillance to maximise knowledge gained on the actual effects of differing adaptations on animal use.

iv

1. Introduction 1.1. Over the past five decades collisions of motor vehicles with deer have escalated in most countries across Europe as well as North America (Ueckermann, 1964; 1987; Lehnert, Romin & Bissonette 1996, Groot-Bruinderink & Hazebroek 1996, Seiler, 2004). Most recent reviews on the subject concur that the toll of deer-vehicle collisions (DVCs) is now likely to be near 1.0 million per annum for Europe (Langbein et al. 2010a) and over 1.5 million in North America (State Farm Insurance, 2009; Mastro et al. 2008). The key factors that have contributed to that rapid growth, not only of DVC but of wildlife road casualties in general, are the expansion of road infrastructure and traffic volumes (Iuell et al. 2003). In Britain for example, traffic volumes have increased fourfold between 1960 and 2000 (Dft, 2006). Still further increases in traffic volume are now showing signs of slowing in some developed countries including Britain (Dft, 2009). In the case of deer, population numbers and distributions are also known to have increased very significantly over the last fifty years in many countries across Europe including Britain as well as in the US (Gill, 1990; Apollonio et al., 2009), adding still further to escalation of DVCs. 1.2. Six species of deer live wild in Britain with a combined population estimated at around 1.5 million (POST, 2009). Following a short-term assessment commissioned by the Highway Agency in 1995 (SGS, 1997), which concluded that DVCs by that time were already very likely to have reached over 20,000 per annum and quite possibly 40,000, a more comprehensive Deer-Vehicle Collisions Project was set up in 2003 through The Deer Initiative. The main aims of that project for the first three years were to assess the true scale and distribution of the problem in England and Scotland, and build a database to help identify hot spots and priority areas for mitigation. Results from the first three years indicated that there are likely to be between 42,500 and 74,000 DVCs each year in Britain (Langbein & Putman, 2006; Langbein, 2007a; Deer Initiative, 2007). 1.3. This high toll of DVCs in Britain leads to extensive human costs (including a number of fatalities and several hundred people injured each year), damage to thousands of vehicles, and presents a major animal welfare issue (as over a quarter of deer are not killed instantly when hit by vehicles, but are left severely injured at the roadside until dispatch or treatment can be arranged). The number of DVCs leading to human injuries in Britain has been assessed to exceed 425 per year including 10 - 20 fatalities (Deer Initiative, 2007). The annual economic „value of prevention‟1 of that level of personal injury road accidents alone amounts to GBP 30M (Deer Initiative, 2007), while additional economic losses in the region of GBP 17M result from the 11,000 or more private vehicles plus 2500 commercial vehicles estimated to sustain significant damage (above insurance excess) from DVCs each year. Further substantial economic consequences of DVCs, for which no national cost estimates are available at present, arise through traffic delays, attendance by suitably qualified personnel to treat or humanely dispatch deer at the roadside and clearance of animal carcasses from the carriageway. 1.4. Although the total numbers of deer in England and Scotland are estimated to be of comparable magnitude (~600,000 - 750,000 each), over 80% of DVCs recorded for Britain as a whole occur in England (Langbein, 2007a). This reflects foremost that English roads make up around 78% of the total road network in term of road length and carry over 85% of annual traffic (Annual traffic reports, DfT 2009), with the inevitable consequence that deer (as well as other species of wildlife) in England are exposed to much greater risk overall of being hit by motor vehicles. Within England DVCs are also not evenly distributed but occur mainly in “hotspots”, often though not exclusively where roads run through woodland with high deer numbers, high traffic volumes and high traffic speed. The most prominent localised hotspots tend to involve fallow, although collisions with roe deer are the most widespread. Overall the species for which highest

1

[Guidance for assessing economic impact of road accidents : Highways Economic Note 1, HMSO 2007]

1

numbers of DVC are recorded in England are fallow (40%), roe (32%) and muntjac (25%); and in Scotland, roe (69%) and red deer (25%). 1.5. Of over 10,000 DVC records accumulated during 2003-2005 for which road type is known, 63% occurred on major roads (A-road and motorways) and 37% on minor roads. Although major roads make up only 12% of the total road length in England they do carry 64% of total traffic volume. The Highways Agency are responsible only for the core strategic road network of motorways and major all-purpose trunk roads in England, which makes up just 1/6th of all major roads (or 2% of roads overall), but which carry close to 33% of all traffic. The precise number of actual collisions on the HA‟s trunk road network is not known. However, over 1100 such incidents have been reported in each of the last three years by trunk road managing agents and Highways Agency (HA) Traffic Officers, and actual numbers on the HA network are likely to be rather higher. 1.6. Estimates based on data accrued by the DI DVC project between 2000 and 2007 indicate that incidents with deer on trunk roads (which make up only around 2% of all road length) are likely to contribute to between 4 - 5% of all DVCs in England. Although the proportion of damage only DVCs tends to be somewhat lower on trunk roads, the percentage of all reported DVCs that lead to human injuries and fatalities is broadly similar for trunk and non-trunk roads in England. We may therefore reasonably estimate, in line with the national economic costs outlined in section 1.3 above, that DVCs on trunk roads alone are likely to incur costs of 2 Million GBP annually through human injuries and damage to vehicles alone. Significant additional impact is, however, likely to arise from DVCs and other encroachment of live deer onto the carriageway especially on trunk roads, because of the high potential here for such incidents to cause serious multiple accidents and traffic delays, as well as the need to deal safely with the humane dispatch of injured deer that are not killed outright in collisions. 1.7. The most successful mitigation measures will seek not to prevent crossings altogether, but to displace them so that deer cross the road in places where accident risk is reduced through lower traffic volume/speed, enhanced visibility and driver awareness, or by provision of relatively traffic free wildlife passages/bridges. Although numerous differing methods have been suggested and tried over the years, the only well proven method of reducing deer collisions on major roads remains the use of roadside fencing of specifications appropriate to the target deer species. However, fencing long lengths of roadway is likely to prove ineffective and result in animals forcing such barrier (with the added risk that they may then become trapped within the carriageway) unless some suitable alternative crossing places are available. 1.8. Wide landscape bridges and other structures constructed specifically for deer and other wildlife offer greatest potential to enable safer crossings and reduce animal collisions; that is, alongside provision of other biodiversity benefits which tend to be their primary purpose (Iuell et al. 2003). Specifications for such wide purpose-built structures are often very high (recommended designs often including width to length ratios above 1.0). The scale and cost of such structures also tends to make them more suitable for new-build roads with multispecies mitigation needs, rather than for widespread application across the existing trunk road network. Deer and other wild large mammals are also known to make use of some motorway and trunk road accommodation structures of much more modest dimensions, that were not built specifically or primarily for wildlife, ranging from footpath and farm accommodation overpasses and underpasses, to viaducts, as well as bridges carrying normal road traffic (e.g. Ballon, 1985; Olbrich, 1984, Langbein, 1996, 2008; SETRA, 1993; Halcrow, 2002). However, the extent and types of existing joint-use structures on the HA‟s trunk road network that may be used by deer in the UK, and the potential for adaptation (of the structure or surroundings) to increase such usage, has received only limited investigation to date. 1.9. This pilot study was initiated to begin to gather information on a range of existing structures on the A38 and A30 trunk roads in Devon and Cornwall, in order to assess whether any are 2

already likely to be used by deer; or else, whether they offer real potential for adaptation in ways likely to bring significant benefits in terms of reducing risk of deer and other wildlife collisions on trunk roads. Requirements for research and objectives of present study 1.10. The majority of recent research into use by large mammals of structures across major roads in Europe and elsewhere has focused on structures either specifically designed for wildlife, or to reduce habitat fragmentation due to road infrastructure (see reviews by Holzgang et al. 2000; Iuell et al. 2003; Georgii et al. 2007). By contrast, studies that have looked into use by wildlife of joint-use structures or ones not created with wildlife in mind, have concerned themselves almost exclusively with structures that also have deer-fencing on both sides of the carriageway, which helps to guide animals to the structures (e.g. Olbrich, 1984, Ballon, 1985m, Langbein, 2008). These and other previous studies addressing the latter issue will be reviewed in further detail in Section 2. 1.11. To date, no systematic assessments have been undertaken for trunk roads in England to assess which types of structures are used or have most potential for use by deer, nor of the factors (including dimensions, locations and surrounding habitats) that influence such use by deer. In particular investigations into the use by deer or by other large mammals of structures without deer fencing are rare, as is knowledge on the extent to which use of such structures in unfenced areas can be increased through provision of deer fencing, habitat manipulation or other adaptations. Objectives for this pilot study 1.12. The main objectives set for the pilot study were as follows:  review past published and unpublished literature (esp. from within UK) on use by deer of structures not specifically designed for wildlife.  to survey and record the characteristics (dimensions / substrate / juxtaposition in relation to differing habitats / deer species) of a large sample of up to 50 existing overpasses, underpasses or viaducts beneath along one or more trunk roads.  to identify if possible any structures with clear evidence of deer use, and / or others with nearby deer activity but no apparent current crossing activity.  assess the potential for further enhancement of use by deer of structures that already appear to be used to some extent, and/or adaptations that might encourage the use of structures which are not currently known to be crossed by deer.  make recommendations for further research, and practical works to enhance use of existing structures by deer on the HA trunk road network.

3

2. Previous research into use by deer of wildlife specific crossing structures and of structures primarily designed for road traffic or other purposes. 2.1. Expanding road networks during the 20th century, and in particular of wide motorways and other major strategic traffic routes, have led to natural habitats and animal populations becoming increasingly fragmented across much of Europe, including Britain. Increased rates of animal mortality through collisions with traffic are another inevitable consequence of that expansion. To help redress the impact of fragmentation through infrastructure on wildlife, in many countries increasing use is being made of green bridges and other means to reconnect habitats and aid animals crossing major roads. Green bridges purpose built for wildlife can also provide optimal solutions for creating safer passage of large mammals across major roads, but the generally high cost of such large structures (commonly upward of 1 to 3 million EU; Georgii et al, 2007) tends to restrict their use to a fraction of the most important conservation sites. The literature and research on green bridges and other means of reducing fragmentation due to infrastructure has been extensively reviewed by others and will not be repeated in detail here, but can be found in the various outputs of the transeuropean collaborative COST341 Action: Habitat fragmentation due to transport infrastructure; in particular see Iuell et. al. (2003), Trocmé et al. (2003) and others available at the IENE web-site). 2.2. The above and other recent reviews of the effectiveness of green bridges (Iuell et al. 2003, Georgii et al. 2007) recommend that to maximise their use by deer, green bridges should ideally be over 40 m wide or else no less than 20 m at their narrowest points, though ideally still with wider entrances. While such demanding criteria would rule out use by deer of the vast majority of existing structures on the trunk road network in Britain, it must be borne in mind that the aims for green bridges tend to be much broader than only to reduce animal mortality and related traffic collisions; among multiple objectives their primary aim tends to be to reduce fragmentation by providing a degree of habitat continuity to encourage use by a high proportion of species and individuals within a given wildlife population. 2.3. The majority of previous research and reviews about the use by deer of smaller crossing structures has also been focussed mostly on assessment of structures specifically designed as wildlife overpasses or underpasses. Several authors have attempted to define minimum criteria for wildlife specific structures, but the minimum dimensions set tend to vary widely (CTGREF, 1978; SETRA, 1993; Ballon, 1985; and see review by Holzgang et al., 2000). CTGREF (1978) analysed the use by large mammals of a range 17 existing overpasses and underpasses that had been built specifically for wildlife on motorways in France, of which only 3 could be positively shown to be used by ungulates. A wider questionnaire study by CTGREF of game managers with regard to 152 different structures, however, did indicate likely use by deer or other ungulates (e.g. wild boar) of 30 wildlife overpasses and 23 underpasses. They concluded, as did Ueckermann (1964) in Germany, that wildlife fencing of appropriate size generally needs to be provided to deter animals from crossing motorway carriageways, and to lead animals instead to suitable crossing structures. 2.4. CTGREF provided some of the first recommendations regarding dimensions for wildlife specific rather than joint-use passages. They suggest a width of 6 m as minimum for overpasses for ungulates, and a height of at least 1/10 of the length or else no less than 3 m for underpasses. Ballon (1985) came to similar conclusions for overpasses, but suggested that width of underpasses should be no less than 8 m for deer. These as well as other early experiences from assessments in France are reviewed by SETRA (1993), to provide various design guidelines for building wildlife specific structures, including that the minimum width of overpasses should be 12 m for red deer, and 7 m for roe deer. Minimum heights for wildlife underpasses are given as 4 m for red deer and 3 m for roe deer, and 2.5 m for wild boar. Berthoud et al. (2000), for wildlife underpasses beneath motorways in Switzerland to be suitable for ungulates, also still recommend a width of 8 m and a 4

minimum height of 3 m provided the passage is no longer than 16 m long, but rising to 4 x 9 m if 25 m long and 5 x 12 m if 32 to 40 m long. 2.5. For „narrow‟ wildlife overpasses on Swiss motorways suitable for all deer Berthoud et al. suggest these need to be at least 25 m wide, thus in effect almost as large as many purpose-built green bridges. However, as discussed above, use by roe deer of much narrower overpasses from around 6 m wide has been reported in some instances, and more recently Langbein (2007c) recorded use also by fallow deer of two farm accommodation overpasses only 4 m and 7 m wide crossing over the M25 in London. 2.6. During research again restricted to assessments of wildlife passages free of any public road traffic, Georgii et al. (2007) by means of video recording and track analysis found roe deer to make fairly regular use (between 0.3 to 5 crossings per day) of six underpasses beneath German motorways ranging in height from just 2.4 to 8.0 m; no larger species of deer were recorded in these underpasses. These authors nevertheless recommended that the height of underpasses designed for wildlife should not be less than 8 m, illustrating that while design criteria will generally err on the side of caution, smaller structures may still suffice as mitigation to aid crossings by individuals of some if not all species of deer present locally. A number of wildlife underpasses 3 to 4 m high have also been confirmed as being used at least by small deer in England, including an underpass beneath the M40 used by muntjac and roe deer (Halcrow, 2002), and ones observed to be used by muntjac, but quite possibly also by fallow, beneath the M25 near Epping and A120 near Stansted (Langbein, 1996; 2008). Georgii et al. (2007) also included 20 purpose built green bridges in their study ranging from 20 m to 200 m wide, finding that intensity of use by wildlife including by deer increased with the width of such bridges. As nearly all, including the narrowest green bridge included in their sample, were used to at least some extent by roe deer, no minimum or threshold widths could be defined for deer. 2.7. Far fewer studies have investigated the use by deer of structures not designed for wildlife but built for joint use for farm or forestry accommodation, footpath diversions or indeed general road traffic. One exception here is a seminal work by Olbrich (1984) who reviewed the use and suitability of near 800 structures of various types on the German Autobahn (motorway) network. All of the structures included in that study occurred in deer-fenced sections of road, and before considering his results, it is interesting to trace briefly the history of the quite extensive use of such deer fencing in Germany and consequent development of ideas for wildlife passages. With as much as 30% of the land area of Germany being covered by woodland, high numbers of collisions with deer were already a longstanding issue even during the early 1960s, leading Ueckermann (1964) to first propose the more widespread use of high fencing to protect traffic against collisions with deer. At the same time Ueckermann mentioned that one underpass in a fenced section had been observed to be used by roe deer as well as foxes and hares. Encouraged by those positive reports the German Transport Minister in 1967 set up a major trial of wildlife fencing across 22 sections of motorway with a total length of 125 km (Busch & Kaemer, 1973), which resulted in a significant reduction in accidents with deer and other game animals within those protected sections. The use of deer fencing in on German motorways expanded rapidly. 2.8. In 1979 Olbrich (1984) for his assessment of overpasses and underpasses selected 49 sections of major trunk routes all of which had been fitted some year earlier with deer fencing. In total 824 crossing structures were identified along these routes, for which site visits and local assessments were made for 788. In each case surveys recorded details of the dimensions of the structures, their location, fencing, and evidence of use by deer and other large mammals. Use by deer was determined by searches for tracks (slots) and dung pellets, or else consultation of each local hunter or forester responsible for deer management in the district where each structure was located. To validate results obtained through views of local deer managers, those structures for which use by large mammals had been reported were visited again after first snowfall in winter to obtain confirmation from assessment of tracks leading across each structure. For roe deer use of 5

structures was confirmed for 45% of underpasses and 22.5% of overpasses located in areas with roe deer present (n=788); a surprisingly high number considering that roe deer have relatively small home ranges and hence possibly rather lesser need to move between sites. Fallow deer were present in the vicinity of 162 structures, of which they were noted to utilize 26.3% of overpasses and 21.7% of underpasses; whereas for red deer among 162 structures within their range, 8.1% of underpasses and 4.8% of overpasses were used. 2.9. Based on analysis of the dimensions, substrate and usage of the structures investigated Olbrich made a number of key recommendations: minimum measurements for height and width of underpasses for deer and other ungulates should be 4 m, with as a matter of principle the shortest possible structures being preferable. The openness index (or relative narrowness) taken for underpasses as width x height divided by length, is highlighted as of particular importance, with suggestions that this should not be less than 1.5 for red deer and also fallow deer. Olbrich found roe deer to prefer underpasses rather than overpasses, and recommends that underpasses should have an openness index of at least 0.75, He suggests that in favourable positions (e.g. quiet farm roads) roe as well as fallow may also use overpasses, but does not provide guidance on their relative openness or maximum length. Olbrich (1984) however concludes that reinforcement of the base (substrate) of the construction with concrete or tarmac makes them more difficult for ungulate use, while proximity to woodland encourages use by large mammals. However, although it is unclear how many of the structures included in his study were open to general road traffic, Olbrich suggests that all structures that are frequented by public traffic are generally unsuitable as wildlife passages. 2.10. Only very few studies to date have looked more closely into the use by large mammals of crossing structures designed primarily to lead public road traffic over or under trunk roads. Pfister (1997) used infrared video cameras along Swiss motorways to study the extent to which crossing structures built for traffic are used as by wildlife. His findings indicated that here such engineering structures may mitigate the barrier effect for burrowing species such as foxes, weasels and martens, but concluded that other more „timid‟ species such as roe deer or hares were seldom or never observed on structures not specifically built for wildlife, even if found in the immediate vicinity of the structure. These findings in Switzerland, however stand in stark contrast to a number of recent studies in England (Langbein, 2007c, 2008) that have confirmed use not just by small deer such as muntjac, but also by fallow deer of a number of structures built primarily for road traffic, and other smaller structures. 2.11. Langbein (2008) assessed the effectiveness of new deer fencing and use of various engineering structures by deer one to three years post-construction of the new A120 Stansted to Braintree dual carriageway. Numerous joint-use over and underbridges were constructed as part of that A120 scheme to accommodate side road traffic, river courses, and/or farm access and footpath diversions. None of the structures, aside from a number of badger tunnels, were specifically designed for wildlife. Deer fencing was provided on either side for 12 km of the new route between Stansted to Dunmow where high densities of fallow were known to occur, and had resulted in nearly 40 DVC per annum locally on the old A120. Investigations of deer use postconstruction used a combination of searches for indirect signs of deer (including use of sand trackbeds on or near entrance to these structures) and intermittent periods of CCTV surveillance. Results demonstrated that both fallow and muntjac deer did make at least some occasional if not yet regular use of several of the structures within two years of construction. The structures where most conclusive evidence including video of deer crossings were found included public road overbridges at Frogs Hall and Warish Hall (see Photos 1 and 2) used by muntjac and fallow, as well as a high underpass bridging the River Chelmer. For a number of other smaller farm and footpath accommodation structures use by deer was reported as having been seen by local landholders but could not be confirmed by video or other signs during the study.

6

Photos 1 and 2: Examples of structures accommodating public road traffic leading over A120 Essex, known to be frequented by fallow and muntjac deer. Both structures cross over deer fenced sections of trunk road; base substrate for both is of concrete or tarmac throughout the main span. (Langbein, 2008) 2.12. It is interesting to note further that the A120 road bridges used by fallow (above) both had concrete or tarmac substrate without any strip of natural surface, although a natural substrate is widely reported as critical to achieve use by deer by many past studies. Another example of a concrete-only overpass used by fallow occurs over the M25 London orbital motorway at Copthall, Epping Forest. Fallow deer have been filmed crossing this 4 m wide and 85 m long concrete-only bridge (see Section 5: Photos 5 and 6), as well as regular deer movement across a slightly wider 7 m overpass which does have a part soil covered base (Langbein, 2007c). Further examples of joint use structures of modest size in England used by deer, include a roadbridge over the M40 near Oxford (Halcrow, 2002) which was provided with a grass and gravel verge to enoucourage wildlife, and use by red deer of an underpass and a narrow overbridge over the M6 Toll motorway (Cresswell pers. comm). In common with all the above examples these structures on the M40 and M6 also occur within sections of trunk roads provided with deer fencing. 2.13. Several of the literature resources discussed above have recommended minimum criteria or conditions for structures that will increase the likelihood of use by deer, and which should be aimed for when planning new-build wildlife passages. However, the exact conditions that lead to some but not other existing structures to be used by deer remain difficult to define. Some consensus does exist that in deciding how inviting or off-putting any man-made structure may appear to a deer, a factor probably more important than any individual dimension is its relative „openness‟ (see 2.9). In simplistic terms shorter wider structures tend to become more readily used by deer than long narrow structures, given other influencing factors remain equal. However, findings from past studies with regard to other factors such as the substrate, location, amount of disturbance, and lead-in fencing and their relative importance remain very variable. The latter is likely to be influenced also by differences between countries and regions in how deer populations are managed and how accustomed deer are to living close to built up areas and human habitation. 2.14. The above review, and in particular the recent observations of structures used by deer in Britain (2.11 and 2.12) illustrate that structures not specifically designed for use by wildlife, including ones of comparatively modest size, given the right conditions may nevertheless become adopted as passages by deer. Furthermore, there may be significant potential to adapt a small proportion of the many hundreds of existing engineering structures over or under the UK trunk road network, to provide safer passage for deer and other large mammals and help minimise risk to traffic. 2.15. Although the examples given where deer use existing joint-use structures are encouraging, the relatively small number and short-term nature of studies that have looked at this issue in Britain suggests this remains a field worthy of much more widespread and detailed research across the UK trunk road network. 7

Table 1: Example dimensions of selected road crossing structures for which deer use has been confirmed during past studies, and recommendations arising from past reviews for design of overpasses or underpasses suitable for wildlife (for more details see text section 2).

Underpasses (UP) Internal height (m)

Existing structures for which deer use reported

Recommended design characteristics to encourage regular use by deer

2.4 – 8.0 (Georgii et. al, 2007) 3.0 - 4.0 (Halcrow, 2002) 4.0 - 7.0 (Langbein, 2008)

>4.0 (Olbrich, 1984) >3.0 for roe deer (SETRA, 1993) >4.0 red deer (SETRA, 1993) >8.0 (Georgii et al. 2007) >4.0 (Olbrich, 1984) >12.0 for red deer (SETRA, 1993) >7.0 roe deer (SETRA, 1993)

Width (m)

4.0 (Olbrich, 1984, Halcrow, 2002, Langbein,2008 ) 8.0 (Ballon,1985)

Length (m) Openness index (width times height divided by length)

up to 48 (Langbein 2007b, 2008) 0.5 (Langbein, 2008)

Variable depending on height / width for roe deer ratio >0.75 for red deer >1.5 (Olbrich, 1984)

6.0 (CTGREF, 1978) 3.5 – 7.0 (Langbein, 2007b, 2008)

>6.0 (CTGREF, 1978) >7.0 (Olbrich, 1984) >25m (Berthoud et al. 2000)

Overpasses (OP) Accessible width (m)

Length (span) (m) Openness (width divided by length) Purpose-built Green Bridges / Wildlife Overpasses

85 -106 (Langbein, 2007b, 2008) 0.05 - 0.06 (Langbein, 2007b)

ratio >0.1 (CTGREF, 1978)

>40 m or at least 20 m at narrowest point with wider entrances (see reviews Iuell et al. 2003; Georgii et al. 2007).

8

3.

Approach and methods

Study site selection 3.1. Fulfilment of the study objectives (1.11), within the limited resources available and a time frame of less than 10 man-days of field work, required selection of study areas that offer a sizeable sample of existing structures (up to 50) within reasonable proximity of one another. To ensure that results would be widely relevant for other parts of the Highways Agency (HA) trunk road network, sites surveyed also ideally needed to include a mix of areas with both small deer species (roe or muntjac) and larger deer (fallow / red / sika). Availability of good background information on DVCs over several past years was also considered important for site selection, to confirm not only the common presence of deer, but also local potential for reducing such accidents in future should the study indicate that adaptation of specific structures could bring significant benefit. 3.2. i. ii. iii. iv.

Alternative sites considered during planning of the pilot study included: A38 Exeter to Bodmin plus A30 Exeter to Bodmin (HA Area 1); Possible extension of above to include M5 Exeter to Cullompton (HA Area 2). A303 Andover to Sutton Scotney ; plus A34 Sutton Scotney to Newbury (HA Area 3). A31-M27 Ringwood to Southampton (HA Area 3).

3.3. Each of the above offered potential to satisfy most requirements for the study. The first option (HA Area 1) was selected foremost on the basis of the particularly wide range of structures available, ranging from small livestock accommodation underpasses to much larger bridges and viaducts spanning river valleys, and secondly availability of comprehensive background information on deer and other animal casualties via the managing agents and through other previous work by the Deer Initiative. 3.4. Review of route plans provided to us by Enterprise Mouchel (HA Managing Agents Area 1) showed the presence of well over 200 existing structures along the A38 between Exeter to Bodmin (including overbridges, underpasses, and culverts) with a similar additional number along the A30 Exeter to Bodmin. As survey of every structure along each route would not be feasible within the time available, the pilot study was focused on structures within six main sub-sections of road chosen to encompass parts of the A30 & A38 with relatively high numbers of reported DVCs, and to ensure representation of some areas with significant presence of roe, fallow and red deer among the sample. Collection of DVC data for the A38 & A30 did not form a specific objective of the present project, but past records available from the DI‟s National DVC project were updated with most recent records available from HA Area 1 managing agents and other sources [to end 2009] and mapped to help with identification of high priority sites. An overview of the relative distribution of reported DVCs along both trunk roads (2003-2009) is provided in Figure 1. 3.5. National deer distribution data indicates at least some occurrence of roe, red, and fallow in almost all of the 20 differing 10 km grid squares overlapping the A30/A38 between Exeter to Bodmin; muntjac have also been recorded as widely present but so far in only about half the area [Figure 3]. While this information suggests any of these species might be encountered almost anywhere along both routes, information on deer presence from local deer managers and from the (small proportion of) DVC reports for which deer species information is available, indicated that fallow would be the most common species contributing to the DVC hotspots (see Figures 1 and 2) along A38 near Haldon, red deer along the A30 west of Okehampton, with roe deer the predominant (but not only) deer species in most other parts. In addition significant numbers of (feral) wild boar are known to be established within less than 1 km of the A30 near Whitstones east of Tedburn St Mary. Muntjac are also likely to be present in small numbers along the route, but no information or direct evidence of their presence was noted during this study. (For the predominant deer species present near different parts of the routes see also Section 4 -Table 3). 9

3.6 Six main road sections were selected on which to focus the survey of existing structures (see Table 3). These were chosen so as to sample some of the main known DVC hotspots along each route, and to encompass some areas where either fallow or red, as well as roe deer are present in significant numbers (see Figure 1 - 3). Along each of these sections the majority of existing underpasses and overpasses were included for survey provided ready access could be gained. Small stream culverts less than 1.5 m high (and least likely to be used by deer) were however excluded. In addition a number of other individual underpasses and viaducts along other parts of the route were also surveyed, either opportunistically if passing through en route to other field work, or selected to help extent the range of different types of structures visited. Field survey of structures 3.7 In line with Health & Safety regulations for working along HA trunk roads, induction training provided by Enterprise Mouchel (HA MAC agents for Area 1) was undertaken by the author prior to commencement of any survey work. As part of requirements for this, Area 1 Command & Control Centre were informed at start and regularly during each fieldwork day if working alone on the network; although for many structures access from or inside trunk road boundary fencing was not required. 3.8 For the purpose of the present pilot study only a one-off survey visit during late March or April 2010 was possible in the case of most structures. Each site inspection covered the following main aspects:  assessment of the main dimensions of each structure (length, width of carriageway and in the case of underpasses, internal height).  search for indirect evidence of recent deer activity (slots / dung / other signs) within or close to structure entrances, and up to 50 m (in some cases 100 m) along the verge of the passage leading over/under the trunk road and on the outside of any highway boundary fencing present.  type of substrate within structure (i.e. hard: man-made concrete / tarmac ; or natural: mud / grass / river-bed )  type and juxtaposition of habitats abutting either side of the structure (e.g. whether scrub / hedgerow or trees present, and approx distance to nearest wooded area).  digital photographs of structures / structure entrances. Dimensions were in all cases measured as the minimum distance for which a deer would need to remain on or within a structure in order to get across. In case of underpasses this was taken as the length under cover excluding any lead-in open to the sky. For further details of information recorded at each structure see survey form – Appendix I. 3.9 Systematic assessments of actual extent of animal usage for every structure (through for example CCTV or track-bed counts to records footprints over extended periods) was beyond scope of the present study. However, for three of those structures where some clear signs of recent deer activity were noted close-by to structure entrances during initial survey, CCTV video recording equipment was deployed for short periods of from 1 to 3 nights. The aim of this opportunistic filming was to obtain some additional insights as to the behaviour of deer crossing through structures or else their behaviour when near them. Two different video set-ups were trialled for this work: a) a set-up available from a previous project developed for surveillance of deer behaviour when crossing main roads and traffic, with a capability for filming and recording at relatively high quality for periods of 24h or more, but also relatively labour intensive in terms of installation and review of extensive amounts of continuous footage recorded; and b) a small pocket-size trail-cam (Scoutguard SGC550) with a video recording function triggered by passive infra-red sensors, and set to start filming for a period of a few minutes when a moving object passes within 10 m in front of the camera. Digital footage obtained was reviewed on computer, to assess presence and behaviour of any deer or other large mammals. 10

Analysis and interpretation of findings 3.10 Data recorded on survey forms and other field notes were catalogued by structure in a spreadsheet format with cross references to photographic images taken on site. In addition placemark references were created using GoogleEarth to enable ready review of aerial views of land surrounding each structure. 3.11 Dimensions recorded during structure surveys were used in the first instance to compare against minimum values for height and width recommended in the literature and to calculate an „openness factor‟ associated with each structure (for underpasses taken as internal height x width divided by length; for overpasses, width divided by length. The openness values obtained were compared with recommendations in the literature (e.g. Olbrich, 1994; Iuell, 2003) of 0.75 for roe deer and 1.5 for larger deer such as red, fallow or sika (Table 1), so as to help assess which structures might have greatest potential to be utilized by deer based on dimension criteria alone. 3.12 However, because visits were “one-off” it was possible to obtain positive evidence of deer use for only a small minority of sites (possibly partly as the great majority of structures had a hard substrate of either concrete or tarmac on which deer tracks do not show up). As a result, no statistical assessment of an association of dimensions or other features with actual deer use could be undertaken on the basis of the findings from this pilot study. Nevertheless, to allow as objective an assessment as possible of the potential of each structure as a safer passage for deer, the main physical features recorded at each site (e.g. openness factor / substrate within passage / presence and level of road traffic / habitats abutting either side, and accessibility for deer of structure entrance) were allocated qualitative scores dependent on whether each feature at a given site was judged to exert a positive, neutral, or negative effect in terms of encouraging deer to make use of that structure. 3.13 In the case of the openness factor for underpasses the criteria suggested by Olbrich (1984) were used to define cut-off points: to allocate single or double positive scores to those structures with openness values exceeding those recommended for roe and for large deer respectively. A negative score was allocated where openness was less than half (=8 m are generally recommended. Most studies conclude that one of the most important feature of underpasses is how open they appear to an approaching deer, with a recommended openness index (width x height divided by length) of >0.75 for small deer species and >1.5 for large deer (red / fallow). 5.8 Such openness criteria remain a useful guide and are met by a very high proportion of the existing underpasses surveyed on trunks roads in the Southwest (24 of 28 underpasses assessed readily exceeded the minimum openness factor for roe deer, while 14 exceeded the higher 1.5 level recommended for large deer), and very likely throughout the UK trunk route network. However, as with all guidance regarding any individual feature, this needs to be interpreted in the local context, as falling short on any one criterion may not be critical provided other conditions are favourable. For example, Georgii et al. (2007) report regular crossings of roe deer of some underpasses as low as 2.4 m with relative openness of just 0.25 to 0.45, while recent findings in England (e.g. see Langbein 2008) indicate that larger deer such as fallow in some situations may also accept structures with an openness well below the 1.5 level recommended for larger deer. 26

Overpass dimensions 5.9 For joint-use overpasses to be suitable for deer minimum width is recommended by most past studies at around 7 m (for detail see Section 2), though ideally wider for red deer. As discussed above, there are however again examples of some structures as narrow as 4.5 m being well used by fallow in some areas. Among 19 overpasses surveyed in the present study all but five in fact exceeded 7m width while only two farm accommodation structures were narrower than 4.5 m. The length of overpasses surveyed during the present study ranged from just 44 to 106 m, with again examples from elsewhere in England of fallow as well as smaller deer using structures at the upper end of that range. As a matter of principle shorter overpasses are generally more likely to be accepted by deer than longer ones, although location and aspect may be even more important. Substrate and traffic 5.10 The presence of a natural base of grass or soil for at least part of the width is another feature likely to improve the probability of up-take as a passage by a wide range of animals including deer (e.g. Georgi et al, 2007; Olbrich, 1994; review Iuell, 2003). Among existing structures included in the present study only 11 out of 47 had a natural substrate for all or part of their width, and in some cases this maybe one of the feature viable for adaptation. However, several overpasses with concrete and tarmac surfaces throughout their span on the trunk roads in England are known to be well used by fallow deer (Langbein, 2008). This is well illustrated by the sample Video Clip (Appendix IV) showing a fallow buck crossing over the six-lane M25 while traffic is passing beneath. Perhaps not least as deer of several species increasingly inhabit such suburban areas in the UK their timidity towards bases reinforced with concrete or tarmac does appear to be reducing. In another example from the present study, fallow deer dung pellets were regularly noted within the tarmac car park of a service area restaurant at Haldon Hill barely 15 m from the A38 trunk road (see Photo 4274). Whist addition of a grass or soil substrate is nevertheless likely to improve uptake of structures by wildlife, Iuell et al. (2003) also conclude that for large mammals including deer width, length and openness considerations are likely to be of greater importance than the substrate. One drawback of the present study was that in view of the very high numbers of structures (36 of 47) with entirely hard substrates this significantly reduced the chances of being able to obtain evidence of use by deer during the one-off inspection visits. To obtain more conclusive results at hard substrate structures repeat visits using sand-traps, video recording or visits after snow cover would be required to enable tracks to be recorded, though more generally where deer do use such structures it is more likely to be under-recorded by comparison to substrates where tracks show up more readily. Traffic / disturbance 5.11 The majority of existing structures surveyed during the present study do carry some road traffic, but the actual size of road and amount of traffic flow varied very widely between them, ranging from some quiet and remote country lanes with minimal vehicle traffic to others with moderate to occasionally heavy traffic. Several previous studies (see section 2) have asserted that structures open to public road traffic will very rarely be used by deer. This however, runs counter to recent findings in England (Langbein, 2008), where both fallow and muntjac deer are known to make regular use of a number of recently constructed public road bridges subject to low to moderate traffic flows (e.g. see Photos 1 and 2, Section 2). 5.12 It is likely that the presence of vehicle traffic will deter use by deer to some extent, but that the actual characteristics of such traffic (e.g. level and type of traffic, and whether largely confined to just part of the day with only occasional use at night) are of greater importance than any clear distinction between presence or complete absence of vehicular access. In the present study, of 36 structures surveyed that are open to some public traffic, 11 were small lanes with only very low levels of traffic even during the day, which would be highly unlikely to deter deer from using them as a crossing. A number of others among those 15 carrying minor roads with light to moderate traffic, by comparison with structures open to motorised traffic used by deer elsewhere in England, 27

would also seem to have some potential for use by deer. On the other hand some structures which carry medium to occasionally high levels would seem less likely to have high potential to be used by deer, and adaptations to encourage such use might in any case not produce significantly „safer‟ crossings in such cases. Habitat 5.13 The habitat adjacent and leading to structure entrances plays an important role in providing cover for deer approaching structures, and indeed increases the likelihood that deer will use areas close to the structure as part of their regular home range. For structures assessed during the present pilot study, nearly half of all the structures had at least some tree or woodland cover adjoining within 50 m on one or both sides, while most of the remainder also had at least some linear scrub cover running along the trunk road verges leading up to the structures. Therefore habitat conditions were not felt to be particularly limiting with regard to possible use by deer, although in many cases improvements would be possible through increasing cover near structure entrance to try and enhance probability of use by wildlife. As with many of the features discussed above, while having woodland cover as close to the entrance as possible would increase probability of use by deer, its presence is not necessarily critical provided location and other local conditions (such as lead-in boundary fencing / high deer density) are favourable. Accessibility 5.14 The accessibility of structure entrances (i.e. how readily deer are able to gain direct access into underpasses or onto overpasses without diversion or significant obstacles), was one of the most variable factors noted during this pilot survey of existing structures. Most common were unfavourable fencing lay-outs (including often lengths of mesh stock-fencing overgrown with hedgerows making them difficult for deer to penetrate) that artificially extend the length of crossing for the passage. Other obstructions noted ranged from gates across entrances of farm accommodation bridges, lorries or horseboxes parked long-term on or within a number of structures, silage bales blocking exits, and areas of hard standing to either side of the structures. Among the samples of structures surveyed accessibility was considered favourable more often for overpasses (12 of 19 sites) than underpasses (14 of 28 sites). In part this difference is explained by fencing for overpasses more commonly being laid out roughly parallel to the trunk road in line with highway boundary fencing. By contrast, roads or tracks leading into underpasses often have additional fencing or thick hedges running on either side, which unless readily crossed by deer may divert them away from the structure entrance. However, in many such cases it is likely that improved accessibility to deer could be achieved through small, site-specific adaptations of the boundary fencing, hedgerow or by provision of small deer leaps [see 4. 5.15 Deer fencing is not currently provided along the trunk road boundary in any of the sections of A38 and A30 encompassed during this pilot study. However, almost all experiences and recommendations arising with regard to wildlife crossing structures in Europe, and also the small numbers of previous assessments in England, are derived from studies of deer fenced roads. It is widely accepted that deer fencing combined with suitable crossing structures can substantially reduce numbers of deer road casualties. Whether or not long lengths of full height deer fencing (which themselves can have negative as well as positive effects) are in fact essential in order to guide deer to use underpasses or overpasses remains a topic for future research. For example no good information is currently available on whether deer can successfully be guided to use existing structures and reduce crossings across the main road by use of only relatively short length of fencing either side of the structures; or whether using lower than full height deer fencing (< 1.8 m) which may not be fully effective at excluding deer, may nevertheless suffice to lead them to easier crossing places. Where wildlife and in particular deer fencing is considered, involvement in the design of the lay-out and installation in the field of an ecologist familiar with deer behaviour is often as important, as the fence specification in terms of height and mesh size. Fencing is often one of the mostly costly parts of a mitigation scheme, but nevertheless poorly installed and inadequately maintained fencing penetrable by deer (often due to underestimation of the gaps that deer can and will exploit) can cause more problems than it solves. 28

Combined potential 5.16 It will be clear from the findings outlined above, and examples of structures for which deer use has been recorded by previous studies, that whether and extent to which any particular structure will be used by deer is unlikely to be determined by any one single feature on its own. All of the different features discussed above (minimum dimensions, openness, substrate, traffic and other disturbance, surrounding habitats, and accessibility) are likely to exert some influence on whether a structure will be used by deer, as well as other factors such as the local deer density. 5.17 The summary Table 4 (Section 4) showing which features of each of the structures visited were judged as being favourable or unfavourable from view of potential use by deer, unsurprisingly shows those found to have the greatest numbers of positive features were mostly high river valley bridges or viaducts, that generally preserve much of the natural land beneath them allowing wildlife to continue to pass relatively freely beneath. Most such viaducts along the two roads studied are already likely be used to some extent by deer, though in some instances adaptation or removal of stock fencing and other possible obstacles could help enhance these further as wildlife passages. 5.18 Table 4 also serves to highlight which structures appear to have the lowest potential as wildlife passages. These will include, for example those with negative scores allocated for high levels of use by vehicles and lowest openness indices. Overall, however, while many existing structures surveyed had some features that would not be considered ideal or be at the lower end of criteria that would be recommended if designing a wildlife passage, it is likely that a quite high proportion (probably more than half) could be adapted in ways to attain some utilisation by deer; not least if resources for some highways fencing were available to help channel animals to specific structures near accident hot spots and deter them from crossing the main trunk road nearby. 5.19 Not all deer may adopt crossing structures as part of their normal daily range that are at the lower end of the generally accepted dimensions (Table 1) and other criteria for wildlife passages. The minimum criteria of greatest importance from view of assessing potential for adaptation of existing structures for purpose of reducing traffic risks through deer collisions, should be those that will enable use of the structures by that proportion of the local deer population most determined to cross the major roads. Deer vehicle collisions, for trunk roads in particular, tend to be highly seasonal (Langbein, 2007a; Deer Initiative, 2007), associated with breeding seasons and annual dispersal phases. Therefore structures that may facilitate only occasional use by deer, or use only at certain times of the year, are still likely to offer high potential to reduce deer collisions. Recommendations for follow-up work 5.20 With the above in mind, Table 4 and other results from the present study also serve as a basis for informing and designing future research trials or demonstration projects. To help assess the practicability and effectiveness of modifications to existing structures in terms of increased deer use and reduced DVC risk, it is suggested that follow up work should include: 

One or more practical trials to show whether targeted limited adaptations to structures near deer vehicle collision hot spots can be shown to reduce deer accident risk through free passage under or over the road.



Practical trials should be designed in such a way as to help also with re-evaluation of the validity and relative importance of minimum criteria that need to be met in order to attain use by large as well as small species of deer (dimensions / joint-use by motorised traffic / substrate / fencing and accessibility).



Assessment of whether structures currently judged as having relatively low potential and with little current evidence of use by deer, increased deer use can be achieved in practice through quite limited and cost-effective adaptations (such as removal of physical obstacles hindering present access by wildlife). 29

Two specific initial trial sites are suggested for consideration for some practical adaptation covering four existing structures (two underpasses, and two overpasses) within HA Area 1, which should include some pre and post works monitoring in each case. Both sites are located in major DVC hotspots, involving in one case fallow deer and roe deer (Haldon Hill), and red deer and roe deer in the other (Ebsworthy Woods). Brief descriptions and what works might be involved in each are provided below. More detailed design of practical adaptations and other work required can be prepared on request. All the recommended adaptations to achieve increased deer use at each of the structures are estimated to be of relatively low cost. In some cases, for example, removal of obstacles or modification of existing fencing to improve accessibility by deer is likely to be achievable at costs below £1,000. In other cases adaptations including provision of short lengths (from around 100 m) of lead-in deer fencing on one or more sides of each structure (@ approx. £30/m) would be estimated to have associated project costs of between 5K and 20K. These broad indicative costs exclude any supplementary follow-up research to assess effectiveness of the measures taken. A) Belvedere Cross and Harcombe Cross underpasses beneath A38 Haldon Hill 5.21 The A38 across Haldon Hill between Kennford and Harcombe has been a deer collision accident hot spot for many years. Haldon Forest holds a fallow deer population well in excess of 500 head, as well as large numbers of roe deer and occasionally some red deer. Between 10 to 20 deer casualties are recorded by the managing agents each year within a distance of around 5 km, while many others are likely to go unreported. Two road underpasses beneath the A38 included in the present study occur within the Forest at Belvedere Cross (Photo 3452]) in the northern part of the forest, and another at Harcombe Cross [Photo 3465] south of the Forest. Signs of deer were found within less than 10 to 50 m of entrance to both underpasses, but not clear evidence of current crossing through these structures. However, extensive frequent fallow deer activity (as shown by video recordings made – see Video Clips 1 – 3, Appendix III) was noted directly on top of the overpass within 2 m of the main trunk carriageways. Deer activity was also noted within woodland adjacent to Harcombe Cross underpass.

Photos 3452 and 3465 – Underpasses beneath A38 at Belverdere Cross (left) and Harcombe Cross (right)

5.22 In case of Belvedere Cross underpass it seems very likely that movements of fallow deer through this underpass could be attained (or increased if already occurring to a limited extent) through installation of a number of short length of fencing. The latter would require careful design lay-out to ensure minimal risk of deer entering the inside of the fenced section through end-runs and maximise its use to deter deer from crossing the trunk road. The opportunity may also exist in this area for possible use of less than full-height deer fencing or fencing erected on the slope of existing embankments that would enable deer to cross in one direction but not another. If such adaptation is found practical after closer investigation and resources are available, some systematic pre and after monitoring of deer use and behaviour around the underpass should be included as part of the scheme, ideally using CCTV equipment or else alternative methods alongside continued monitoring also of deer collisions, in order to assess the actual effectiveness and cost effectiveness of adaptations undertaken. 5.23 Harcombe Cross underpass (Photo 3465) is one of the widest and highest road underpasses surveyed, benefitting also from wide footpath / sideways either side of the carriageway carrying intermittent low levels of one-way traffic leading onto the A38. The main 30

adaptations to consider here might be to provide a more natural surface along one of the wide sideways to assess whether possibly increased use by deer can be created through that means alone. Again short length of deer fencing (c.200m) along either side of the wood traversed by the A38 in this area could also be considered to help channel deer towards the underpass. As above (5.18) monitoring of deer use before and after such works should be part of the scheme. B) Ebsworthy Wood (Ebsworthy Farm and Ellacott Farm Overpasses over A30) 5.24 Ebsworthy Woods lies 2 miles from the northwestern boundary of Dartmoor National Park. Deer road casualties are not only a recent problem here but have been recorded in this area for a long time, in particular along a 1 km stretch where the A30 bisects the woodland and where wildlife warning reflectors and signs have been installed for several years in attempt to mitigate the problem. The fact that in recent years red deer (which at around average >100 kg for adult females and > 150 to over 200 kg for adult males, are around twice as heavy as fallow and four times as heavy as roe deer) are crossing the road in this area is of particular concern, as seriousness and likelihood of injury in DVCs is known to increase with size of the species involved. Information on the species of deer involved has mostly been unavailable from past records provided by the managing agent and other sources, and the proportion of DVCs involving red deer in this area is not currently known. (A request to keep species details and/or take a digital image of all animal casualties collected has been made by the EM ecologist to obtain more detailed information in future). The presence of red deer in the area is well known, and was also confirmed during the present survey (e.g. see image Photo 4592). Two overpasses occur 2 km apart to either side of Ebsworthy Wood. Ellacott Farm Overpass and Ebsworthy Farm overpass. 5.25 Ellacott farm overpass (Photos 4111; 4119) is a narrow farm accommodation bridge free of any road traffic and only occasional farm vehicle access. At the time of survey a lorry had been parked continuously on the structure for several months, blocking much of the path (Photo 4111) , whilst a metal gate and piles of silage bales block much entrance onto the structure from the south. Signs of deer were found within less than 50 m on the northern side, though no signs noted directly on the (hard substrate) bridge itself. Although this overpass is only 4m wide, it is very favourably located to serve as a deer crossing in a very quiet area where Ebsworthy wood and a new tree plantation abut closely to its northern side and pasture fields to the south. The main adaptation suggested in this instance would centre around removal of current obstacles to deer use of the structure; i.e. by ensuring no farm vehicles or other machinery are parked long-term on the structure in future, and other obstacles such as silage bales and gates are removed to enable freer passage for deer and other wildlife. Stock fencing adjoining the structure especially on the southern side should also be adapted in a way that enables ready access by deer. In addition, consideration should be given to improvement of the woodland / trunk road boundary fencing for the c. 2 km section located between the two overpasses, to help channel deer to the safer crossing places and prevent them crossing over the trunk carriageways in this area.

Photos 4110 and 4096 – Overpasses over A30 at Ellacott Farm (left) and Ebsworthy Farm (right)

5.26 Ebsworthy Farm overbridge is a somewhat wider (7 m) overpass carrying a minor road / lane (e.g.Photo 4096, 4102) with very light local traffic flow. Woodland abuts within 75m on the southern side, with pasture and moorland to the north. Signs of red deer activity were found within 75m of the entrances, and although no direct evidence of deer crossing activity was noted during the short pilot survey, it seems likely that deer may already use the bridge to some extent. The 31

main adaptations that might be considered to enhance this passage for use by deer would include minor adaptations to the stock fencing surrounding woodland on the south side to enable more direct access by deer from the wood near the overpass entrance. Creation of a soil, grass other natural substrate along one or both existing concrete verges, and some improvements to existing lay-out of lead-in fencing would also be considered beneficial here. In addition (as outlined in 5.24) improvements to existing trunk road boundary fencing between this structure and the second overpass at Ellacott Farm should be considered to maximise use of both structures by the deer. Some repairs to that stretch of fencing (believed to the responsibility of the landowner abutting the HA estate) are required at present in any case as that fencing has been damaged by errant vehicles involved in accidents over the past year. 5.27 If any adaptations of the overpasses to enhance their joint use by wildlife are initiated, then as suggested also for A above (5.21) some systematic pre and after monitoring of deer use and their behaviour on and around the structures (ideally using CCTV) should form part of the scheme. 5.28 The two areas for local adaptation trials outlined under A & B above have been chosen for their potential to serve as demonstration sites of what may be achievable through similar adaptations also in other areas of the network, and in order to test past assumptions regarding the characteristics of suitable structures for use by deer and other large mammals. Many of the other structures surveyed along the A30 & A30 could similarly benefit from various similar though sitespecific adaptations should resources be available for trials at a wider selection of sites. 5.29 Consideration should also be given to a desk-top study of other areas of the HA trunk road network throughout England to help identify a selection of those structures in each HA region likely to offer greatest potential for adaptation to mitigate DVC risks (possibly initially based on use of existing on-line aerial mapping resources and information on animal collision hotspots).

Acknowledgements This study was enabled by a grant from Highways Agency to The Deer Initiative. Majority of fieldwork was undertaken by the author, who would like to acknowledge the input and local information provided by John Stowers and Brian Johnson who each accompanied him for one field survey day, and to Leo Gubert, Ed Dyson, David Hinde and Richard Bernhardt for helpful discussion during site visits and comments on a draft version of this report.

32

References Apollonio, M., Andersen, R., and Putman, R.J. (eds.) (2010a) European Ungulates and their Management in st the 21 century. Cambridge University Press. Ballon, P. (1985) Bilan technique des aménagements réalisés en France pour réduire les impacts des grandes infrastructures linéaires sur les ongulés gibiers. In: XVIIth Congress of the International Union of Game Biologists, Brussels. 679-689. Berthoud, G. et al. (2000) (15+ authors ; see IENE web-site for fuller detail): COST341 Habitat fragmentation due to Tranport Infrastructure. Swiss State of the Art National Report. European Commission, Brussels. Busch, F. and Kraemer, E. (1973) Wildschutz an Bundesstrassen. Strasse und Autobahn 24, 172-173. Cresswell, W. , Dean, M., Jones, R., Langbein, J. and S. Wray (2008) Use of underpasses, culverts and tunnels under roads by mammals. Presentation to The Mammal Society‟s Autumn Symposium,Nov.2008. Institute of Zoology, London. CTGREF (1978): Autoroute et grand gibier, 41 S. (Note technique No. 42). Groupement technique forestier, Nogent-sur-Vernisson, Paris. Deer Initiative (2007) Deer on our Roads – Counting the Costs. Deer Initiative, Wrexham. DfT (2006) Transport Statistics Great Britain (TSGB) 2006 Edition. DfT (2009) Transport Statistics Great Britain (TSGB) 2006 Edition. Department for Transport web-site http://www.dft.gov.uk/ Georgii, B., Peters-Ostenberg, E., Henneberg, M., Herrmann, M., Müller-Stieß H. and Bach, L. (2007) Nutzung von Grünbrücken und anderen Querungsbauwerken durch Säugetiere. Gesamtbericht zum Forschungs-und Entwicklungsvorhabe. [In German] Forschung Strassenbau und Strassenverkehrstechnik Heft 971. Bonn. Gill, R.M.A. (1990) Monitoring the Status of European and North American Cervids. GEMS Information Series, 8; Global Environment Monitoring Systems, United Nations Environment Programme, Nairobi. 277 pages. Groot Bruinderink, G.W.T.A. and Hazebroek, E. (1996). Ungulate Traffic Collisions in Europe. Conservation Biology 10, 1059-67. Harris, S., Morris, P., Wray, S. and Yalden, D., (1995) A review of British mammals:population estimates and conservation status of British mammals other than cetaceans. JNCC, Peterborough. Holzgang, O. , Sieber, U., Heynen, D. , von Lerber, F., Keller, V. and H-P Pfister (2000) Wildtiere und Verkehr: Eine kommentierte Bibliographie. Schweizerische Vogelwarte, Sempach. IENE web-site. www.cbm.slu.se/iene/cost341.php Iuell,B., Bekker, G.J., Cuperus, R. Dufek,J., Fry, G., Hicks,C.,Hlavac, V., Keller, V.B., Rosell, C., Sangwine, T., Torslov, N, Wandall, B. le M, (2003) Wildlife and Traffic: A European Handbook for Identifying Conflicts and Designing Solutions. European Commission Action 341 on "Habitat Fragmentation due to Transportation Infrastructure", Brussels. Halcrow (2002) Managing Integration : Highways Agency -Minimising Fragmentation. Efficacy of Existing Green Bridges M40 between Junctions 8 and 9. Research Report. Halcrow Group Ltd. Swindon. Langbein, J. (1996) Conservation and management of deer in Epping Forest and its Buffer Land Estate. Corporation of London. Langbein, J. and Putman, R. J. (2006). National Deer-Vehicle Collisions Project : Scotland 2003-2005. Final report to Scottish Executive. The Deer Initiative, Wrexham. Langbein, J. (2001) A120 Stansted to Braintree Bypass: Pre-construction Deer Studies. OverArup, Solihull. Langbein, J. (2007a) National Deer-Vehicle Collisions Project : England 2003-2005. Report to The Highway Agency. The Deer Initiative, Wrexham. Langbein J. (2007b) Use of remote video surveillance to investigate deer behaviour in relation to wildlife deterrents, roads and vehicles. Presentation at „Deer on our Roads Seminar‟, Ashridge, UK, October 2007. [http://www.deercollisions.co.uk/pages/workshop2.html] Langbein J. (2007c) Epping Forest Deer Management Review. Report to Open Spaces Department, City of London. Langbein, J. (2008) A120 Stansted to Braintree post-construction deer studies. Balfour Beatty. London. Langbein, J., Putman. R.J., and B. Pokorny (2010b) in press. Road Traffic Collisions involving deer and other ungulates in Europe and available measures for mitigation. In: Ungulate Management in Europe: Problems and Practices (eds. R.J. Putman, M. Apollonio and R. Andersen), Cambridge University Press, in press. Lehnert, M.E. and Bissonette, J.A. (1997). Effectiveness of highways crosswalk structures in reducing deervehicle collisions. Wildlife Society Bulletin 25, 809-818.

33

Mastro,L.L., Conover, M.R. and Frey, S.N. (2008) Deer-vehicle collision prevention techniques. HumanWildlife Conflicts 2, 80-92. Olbrich, P. (1984). Untersuchung der Wirksamkeit von Wildwarnreflektoren und der Eignung von Wilddurchlässen. Zeitschrift für Jagdwissenschaft 30, 101-116. OveArup (1995) A120 Stansted to Braintree Bypass, Environmental Statement. PFISTER, H.P. (1997): Wildtierpassagen an Strassen. Vorprojekt zur Abklärung der Nutzung von für den Verkehr erstellten Unter- und Überführungen durch Wildtiere. Forschungsauftrag 30/92. Bundesamt für Strassenbau und Vereinigung Schweizerischer Strassenfachleute (VSS), Zürich. 29 S. Pfister, H.P., D. Heynen, B. Georgii, V. Keller, and F. von Lerber. (1999). Häufigkeit und Verhalten ausgewählter Wildsäuger auf unterschiedlich breiten Wildtierbrücken (Grünbrücken). Schweizerische Vogelwarte, Sempach, Switzerland. POST (2009) Note number 325 Wild Deer in the UK: Impacts of rising deer populations. The Parliamentary Office of Science and Technology, London. Putman, R.J., Langbein, J. and Staines, B.W. (2004) Deer and Road Traffic Accidents; A Review of Mitigation Measures: Costs and Cost-Effectiveness. Report to the Deer Commission for Scotland. Seiler, A. (2004). Trends and spatial pattern in ungulate-vehicle collisions in Sweden. Wildlife Biology 10, 301-313. SETRA (1993) Passages pour la grande faune, Guide technique. Service d'Etudes Techniques des Routes et Autoroutes, Bagneux. pp. 121. Smith, G. and Langbein, J. (1998) (eds). The Prevention of Wildlife Casualties on roads through deterrents: Prevention of casualties amongst deer populations. Final Report SW335/V3/11-98 to Highways Agency. SGS Environment Ltd. Trocmé, M.; Cahill, S.; de Vries, J.G.; Farrall, H.; Folkeson, L.G.; Hicks, C. and Peymen, J. (eds), 2003. COST 341 - Habitat Fragmentation due to Transportation Infrastructure: The European Review. Office for official publications of the European Communities, Luxembourg. Ückermann, E. (1964) Wildverluste durch den Strassenverkehr und die Verkekrsunfaelle durch Wild. Z. Jagdwiss. 10: 142-168. Ückermann, E. (1983). Die Auswirkung der Wildverluste durch den Stassen Verkehr auf die Nutzungsfähigkeit der Reviere. Zeitschrift für Jagdwissenschaft, 29, 264-265.

34

Appendix I : Field Survey Form

Existing Structures Survey Sheet March/April 2010 3. Site Name:

1: Recorder

2. Date of initial survey:

4.Road:

5. Structure Type:

8 Width:

9. Height

6: General description:

Dimensions (m): 7. Length:

10.Substrate:

Vegetation 11a: Carriagway A

11b: Carriagway B

12. Deer use signs: (position / distance / frequency) 12a Slots 12b Pellets 12c Trackways 13. Other Comments (incl. other animal signs):

14: Photographs taken:

Yes / NO

15. Overall Potential of structure:

Images nos:

15a Score:

Appendix II : Tables 5a & (overleaf) 5b

Table 5a: Summary of main features recorded during field survey of Overpasses Structure Name Brewers Cowsen lane Ebsworthy Farm Barton Head Bible Christian Venn Bridge Hill Hackworthy Lane Marely Head Lanhydrock Merafield Road Deep Lane Jnc Holewell Lane Saltram House Milestone Lane Old Exeter Road Ellacott (Ebsworthy) Forder Lane Horsebrook Lane A30 Halgavor NCN

Road / Str number A30/320.80 A30/308.30 A30/312.20 A30/273.3 A30/369.40 A30/280.60 A30/276.5 A38/75.50 A30/367.70 A38/54.20 A38/57.50 A30/282.30 A38/53.30 A38/103.2 A38/102.7 A30/314.10 A38/69.50 A38/73.40 A30/367a

Type O/B O/B O/B O/B O/B O/B O/B O/B O/B O/B O/B O/B O/B O/B O/B O/B O/B O/B O/B

Length m 46 51 46 64 65 60 64 77 73 101 106 73 58 80 90 44 76 102 59

Width m 9.5 8 7 9.5 9 8 8.5 10.2 9.5 12.5 13 8 6.2 8 8.5 4 6 6.5 3.5

Substrate - Hard (H) or Natural (N) H: concrete / tarmac H: concrete / tarmac H: concrete / tarmac H: concrete / tarmac H: concrete / weedy H: concrete / tarmac H: concrete / tarmac H: concrete / tarmac H: concrete / weedy H: concrete / tarmac H: concrete / tarmac H: concrete / tarmac H: concrete / tarmac H: concrete / tarmac H: concrete / tarmac H: concrete / tarmac H: concrete / tarmac H: concrete / tarmac H: Playground type

Openness Factor 0.207 0.157 0.152 0.148 0.138 0.133 0.133 0.132 0.130 0.124 0.123 0.110 0.107 0.100 0.094 0.091 0.079 0.064 0.059

Poss. Deer signs Yes/inc. Yes Yes Yes Nf Nf yes Nf Nf Nf Yes yes Yes/inc. Nf Nf Yes Nf Yes/inc. Yes

Nearest sign